Gas turbine maintenance and inspection model enhancement

Maintenance, Reliability and Inspection

S. A. ALSHAHRANI, Saudi Aramco, Dhahran, Saudi Arabia

Improving asset performance is a major concern throughout industry and is associated with maintenance costs, experience, machine/equipment age, and operational and environmental conditions. This article will focus on improving the reliability and availability of a plant gas turbine by applying best maintenance practices, operational enhancements and performance monitoring. Moreover, an inclusive maintenance and inspection model is proposed compared with the current common model.

The subject model consists of preventive maintenance and condition monitoring that ensure a cost-effective maintenance philosophy. Quality assurance and quality control (QA/QC) are key elements in the subject model as they will measure all steps and scope and provide an improvement or change to the system processes or practices. Additionally, fault tree analysis (FTA) and reliability block diagrams (RBDs) were the main tools used to identify major challenges on gas turbine reliability and availability.  

Current gas turbine maintenance and inspection model. The current maintenance and inspection model of the gas turbine is based on manufacturer guidelines and manuals. The strategy followed is time-based maintenance (scheduled maintenance): after a certain number of hours in operation, the machine is stopped and inspected, regardless of its current condition. However, if the operation or maintenance team notice any abnormal condition, the machine is stopped and checked accordingly, and a necessary unplanned inspection might be performed. FIG. 1 demonstrates the current model including the main inspection cycles: 

• Combustion inspection (CI): A CI is the inspection of the gas turbine parts that are inside the rubber casing located between the axial compressor and the 1st stage turbine blade. Generally, the inspection is performed every 8,000–12,000 running hours (hr), which is approximately equivalent to 1 yr–1.5 yr depending on the operational needs and number of the critical/ technical issues discovered. A CI is performed twice before any major inspection (MI), as indicated on the model. 

• Hot gas path inspection (HGPI): This inspection includes an inspection of combustion parts inspection as well as turbine blades. This inspection is performed after no more than 24,000 running hr from the last MI—between the two CIs. At this stage, the turbine casing cover is removed so the turbine blades of both rotors are exposed for visual inspection and replacement, if needed. In some gas turbines operated with fuel oil, the turbine blades will be replaced due to accumulated deposits and carbon as result of the firing temperature and combustion system. 

• MI: This calls for an inspection from flange to flange, entailing the entire gas turbine from the bell moth (at Bearing 1) to the end of the exhaust (at Bearing 4) in most gas turbine designs. This inspection is conducted after 48,000 hr—equivalent to 5.5 yr—and follows two CIs and one HGPI. The MI scope is huge and requires significant resources and planning. Every part will be exposed, checked and replaced, as recommended.  

Current model limitations and case study. It was noticed that the performance of the gas turbines dropped even after applying and implementing manufacturer recommendations (the current model). Premature failures for turbine internal parts discovered during the inspection included: 

• Bearing failures 

• Oil leak at Bearing 1 

• High exhaust temperature 

• Axial air compressor blade damaged  

• Wear of internals parts  

• Performance drop  

• Issues with auxiliary systems.  

These failures were the result of inefficient scope, poor operator skills, the absence of effective QA/QC, and the low quality of the refurbished parts used in the overhaul. Each part of the gas turbine has a specific lifetime and should not be repaired and reused after that limit. Unfortunately, there is no mechanism in plants or repair shops to effectively track each part’s running hours, particularly for parts that are removed (e.g., liners, transition pieces, fuel nozzle, cans, buckets). 

According to the history of the scheduled maintenance for the last 20 yr, no issues or findings on the turbine blades were found during an HGPI. This indicates that the HGPI scope can be optimized and replaced with a CI scope. Also, a borescope device can be utilized to check the blade conditions instead of removing the turbine casing, which will lead to improved utilization of resources.    

In addition, maintenance costs increased due to inflation and the replacement of gas turbine internal parts. The loss of experience in gas turbine maintenance personnel also contributed to high maintenance costs and decreased machine performance.  

A lack of clear inspection guidelines caused the maintenance team to change out all gas turbine internal parts without assessing their condition, which would prevent reoccurring failures. 

As an example Case Study, it was observed during an MI that damage to Bearing 2 was not discovered until the machine was opened after 48,000 running hr because there is no condition monitoring program on the subject bearing.  

For further improvement of gas turbine reliability and availability, the fault tree analysis (FTA) methodology and reliability block diagrams (RBDs) are used and discussed in the following sections. A failure mode and effect analysis can also be used to determine system challenges or other issues related to the gas turbine, but is not part of this discussion.  

FTA. This methodology is effective in finding the root case analysis of the issues and determining the relationship between the top event and contributing challenges, either direct or indirect. FIGS. 2A–2F represent 24 major challenges related to the low reliability of gas turbines, including poor performance and high maintenance costs. Challenges were classified into two types that must be addressed and mitigated: direct and indirect challenges. FTA depends on two gates, the AND gate and the OR gate. An AND gate indicates that the top event will happen if the both direct and indirect challenges happen, whereas the OR gate indicates that any challenge will result to the top event.

The major issues applicable to each major challenge are summarized below.  

Mechanical issues. Vibration issues can go undetected (hidden failure) as there are no designed vibration measurements at the internal bearings. When the bearings were removed during the MI, it was observed that the bearing clearances were high and there were scoring marks on the shaft.  

• Signs of high bearing temperature were noticed on the bearings after removal. This was observed again on internal Bearings 2 and 3.  

• High exhaust temperature was discovered, which is a major issue related to unreliable fuel nozzles. Sometimes fuel nozzles are blocked with carbon deposits, or they inject more fuel than they should.  

• High compartment temperature was observed with leakages of the compartment’s insulation. Also, the ventilation was less than adequate.  

• Wear of internal parts was noticed during the inspection of the gas turbine. Two main reasons are inaccurate installation and internal vibration (shaking) while the machine was in operation.  

• Oil leak was found at Bearing 1; this is a common issue that is related to loss of air seal as a result of blockages or dust that passes the air intake filter.  

Electrical issues. These include: 

• Loss of power occurred only once due to loss of the backup ultranet power supply (UPS) while preventive maintenance was being performed on the primary UPS.  

• Cables were occasionally found to be grounded due to their age and infrequent integrity checks.  

Instrument issues. These include: 

• Control valves fluctuate frequently if one of the parameters exceeds its range. Depending on the valve and its age, frequent calibration and tuning are required.  

• Calibration issues are similar to—or part of—the issue above, depending on personnel experience and equipment condition.  

• Spike readings can happen suddenly without previous indication. Spike readings trip the machine, affect overall gas turbine performance and push the part to its fatigue cycle.  

• Control cards failures occur frequently and are related to high temperatures inside the control room.  

Maintenance issues. These include:  

• Improper inspection of internal parts, as most parts during the inspection are sent to the local manufacturer shop for refurbishment without detailed inspection. 

• Improper installation was noticed: there was wear on one side between the contact areas (e.g., liner springs) and contact areas on transition pieces.  

• Current QA is ineffective or nonexistent. QC checks repair quality (e.g., internal clearances) but is not comprehensive.   

• Root cause analysis (RCA) is ineffective as no comprehensive report discusses all findings during the inspection.    

Operational issues. These include: 

• Frequent trips during startup shorten the gas turbine lifecycle by affecting hot path parts.  

• Daily monitoring is not performed. Alarm rationalization is not part of the operational reliability focus due to excessive daily noise alarms.  

• Operators lack experience for when machines are operated beyond their operating windows. 

• No technical courses are available for personnel. Operators should have the minimum of knowledge and work to improve their skills. Machine reliability is everyone’s responsibility.   

Machine aging. This involves: 

• Machines aging is a common factor throughout industry and exacerbates the above causes or conditions.  

RBD. RBDs are a representation of the FTA relations seen in FIG. 3. The relation of a top event (gas turbine challenges) are parallel due to the AND gate between the direct and indirect challenges. The relationship between all other issues is due to the OR gate. Each event is demonstrated by box and number. The events from 1–13 are related to direct challenges, while the events from 14–24 are related to indirect challenges.

For example, the minimum cut set is used to determine the possible causes of the top event, then the low performance of the gas turbine could be due to Events 7 and 19, which are low speed and an ineffective RCA, respectively. Also, the cause could be related to 1.14, 1.15, 1.16…and so on. In total, there will be 143 related issues if considering the RBD in FIG. 3. Therefore, both direct and indirect challenges contributed to the top event.

To improve and enhance gas turbine reliability and availability, it is highly recommended to implement the following: 

• Apply a condition monitoring program rather than a “run-to-fail” strategy, especially for air intake filters and vibration monitoring.  

• Conduct reliability awareness sessions for operations and maintenance personnel.  

• Place an emphasis on the root cause practice to be performed for all failure. 

• Conduct failure mode and effects analysis (FMEA) or any other equivalent methods to analyze the system.

Proposed model of gas turbine inspection and maintenance. Because following manufacturer recommendations did not yield high-quality and low-cost maintenance, current practices and manufacturer recommendations must be challenged. The current model should be improved in a way to capture most of the issues in the current model and maintenance practices as identified in the FTA and RBD.

The objective of the proposed model is to extend the lifetime of the machine—as well as increase inspection intervals—with maximum throughput and efficiency. Therefore, the model focuses on three main areas: pre-maintenance and inspection, maintenance and inspection, and improvement (as seen in FIG. 4). Within each of these areas, several steps are needed to reach the ultimate goal and objective. This comprehensive model should mitigate the current model issues in planning, execution and tracking, and ensure all subject matter experts (SMEs) are participating efficiently and effectively. For example, the current model planning might be performed by one entity from the turnaround group or sometimes only by the planning engineer. However, the comprehensive model is most effective when all experts are part of the process.

Pre-maintenance and inspection include all inputs needed to execute the work on schedule and on (or under) budget. At this stage, documentation, planning and QC must be performed effectively. Documentation includes all paperwork that will be used as a reference to review old inspections and execute the current work with high quality. Planning is essential to organize the work and should dictate the what, when, how and where of the work. QC maintains the compliance of all inputs with respect to standards, scope, requirements, key performance indicators (KPIs), etc. before the job is executed.

The maintenance and inspection phase comprises schedule maintenance and condition monitoring maintenance. This phase presents some significant changes compared to the current methods of inspection: 

• Eliminating the HGPI in accordance with the machine’s history 

• Extending the interval between the inspections to be up to 15,000 hr rather than 12,000 hr based on the condition review of the previous inspections 

• Conducting initial inspection for the removed internal parts 

• Introducing the condition monitoring as an essential element, especially for air intake and vibration monitoring systems.  

The improvement phase consists of two main steps: QC and reporting and recommendation implementation (R&RI). QC checks the activities executed against the planned scope, schedule, cost, standards, drawings, safety, etc. QC is ongoing and must be conducted regularly—for some tasks, this means daily or even every shift. R&RI is the final stage where all findings, RCA and recommendations are recorded. Additionally, KPIs are measured to check performance, compliance, efficiency or other measurements.    

Modified FTA and RBD. After implementing the new proposed model, the expected FTA is represented in FIG. 5. Failures or malfunctions cannot be eliminated, but they can be predicted or minimized in terms of occurrence and frequency. For example, in the FTA, mechanical failures can be avoided only when condition monitoring is implemented. Also, gas turbine challenges will occur if both direct and indirect challenges have an issue or a failure occurs. Moreover, the RBD in FIG. 6 is plotted to determine the system boundary and most-effective shortcut. It is clear that the minimum shortcut could be 1.5.9.10, 2.5.9.10, 3.5.9.10 or 4.5.9.10. This means that high vibration will be a challenge without a condition monitoring program, improper operation and no daily operation rout.

 

In addition, it is expected that system reliability will be improved further compared with the old model due to the addition of AND gates, which indicate further redundant protection programs.

Takeaways. In any industry, maintenance is a major concern and can consume a significant portion of an organization’s resources. Each maintenance philosophy must be reviewed and enhanced to ensure successful implementation within the allocated budget. The proposed model can be implemented and customized to any maintenance activities to enhance overall plant and asset availability and reliability. If the proposed model is implemented, the following benefits can be expected:  

• Minimize the gap in the scope as the model is comprehensive and includes two main gates (QA/QC).  

• Reduce the number of trips and failures, which means a reduction in the thermal effect on the gas turbine combustion component.  

• A drop in maintenance costs because of a maintenance strategy change to more proactive maintenance rather than reactive maintenance.  

• An associated risk will be minimized by addressing the root cause of the internal failure.  

• The level of experience will be improved due to the additional onsite scope and examination of the removed parts.  

• Demand will be meet with high performance as monitored by proposed KPIs. GP&LNG

ABOUT THE AUTHOR

Saeed Ali Alshahrani is a Maintenance Engineer Specialist with Saudi Aramco and has more than 16 yr of experience in the maintenance segment. He has extensive experience in rotating equipment operation, maintenance and analysis, and is familiar with maintenance indicators (KPIs) that support asset improvement and overall plant availability and reliability. He earned a degree in mechanical engineering from KFUPM and a master’s degree from Manchester University in reliability engineering and asset management.

Related Articles